[fnovd]

>affordable universal EMR would significantly enhance the ability to deliver care for almost all sizes of healthcare providers.

Additionally, it would open the door for EHR-based research on a massive scale. Getting different hospitals' EMRs to play nice with one another is a huge barrier for large, retrospective cohort studies. Projects like the Million Veteran Program (http://www.research.va.gov/mvp/) would be just the beginning.

[darkmarmot]

I work in this space in the private sector. From my perspective, the greatest impediment in the U.S. relates to our laws regarding PHI. I work on a system with petabytes of medical data -- but researchers can't access it because our laws and regulations don't support it (and thus provide no financial incentive and only risk and lawsuits for the companies holding it.) It's abysmally sad.

[Eugr]

Well, PHI protection (HIPAA) is actually a good thing and doesn't really prevent researchers from accessing data as long as PHI is stripped (de-identified dataset). Now getting that de-identified data is not always easy...

[gmarx]

Now you've hit my area of expertise. This might help enable large retrospective cohort studies but would not be sufficient. In order to do that the universal system would need to do several things which are unlikely: 1)Universal clinically meaningful data model 2) Extensive use of standard terminology codes 3) good model of timeline 4) user interface which encourages the entry of discrete data over free text